Examples with MultiFieldQueryParser org.apache.lucene.queryparser.classic.MultiFieldQueryParser used on opensource projects

Example 21 with MultiFieldQueryParser

use of org.apache.lucene.queryparser.classic.MultiFieldQueryParser in project Anserini by castorini.

the class EntityLinking method search.

/**
 * Returns a list of query results.
 *
 * @param queryName the entity name to search
 * @throws Exception on error
 * @return a list of top ranked entities
 */
public List<RankedEntity> search(String queryName, int numHits) throws Exception {
    List<RankedEntity> rankedEntities = new ArrayList<>();
    // Initialize index searcher
    IndexSearcher searcher = new IndexSearcher(reader);
    // do exact search on query name
    QueryParser queryParser = new QueryParser(IndexTopics.FIELD_NAME, new SimpleAnalyzer());
    queryParser.setAutoGeneratePhraseQueries(true);
    queryParser.setPhraseSlop(3);
    queryName = "\"" + queryName + "\"";
    Query query = queryParser.parse(queryName);
    TopDocs rs = searcher.search(query, numHits);
    ScoredDocuments docs = ScoredDocuments.fromTopDocs(rs, searcher);
    for (int i = 0; i < docs.documents.length; i++) {
        float score = docs.scores[i];
        String mid = docs.documents[i].getField(IndexTopics.FIELD_TOPIC_MID).stringValue();
        String shortMid = getShortMid(mid);
        String name = docs.documents[i].getField(IndexTopics.FIELD_NAME).stringValue();
        String label = docs.documents[i].getField(IndexTopics.FIELD_LABEL).stringValue();
        rankedEntities.add(new RankedEntity(shortMid, score, name, label));
    }
    if (docs.documents.length >= numHits) {
        return rankedEntities;
    }
    int numHitsLeft = numHits - docs.documents.length;
    // do TFIDF search
    Similarity similarity = new ClassicSimilarity();
    searcher.setSimilarity(similarity);
    queryParser = new MultiFieldQueryParser(new String[] { IndexTopics.FIELD_NAME, IndexTopics.FIELD_LABEL }, new SimpleAnalyzer());
    queryParser.setDefaultOperator(QueryParser.Operator.AND);
    query = queryParser.parse(queryName);
    rs = searcher.search(query, numHitsLeft);
    docs = ScoredDocuments.fromTopDocs(rs, searcher);
    for (int i = 0; i < docs.documents.length; i++) {
        float score = docs.scores[i];
        String mid = docs.documents[i].getField(IndexTopics.FIELD_TOPIC_MID).stringValue();
        String shortMid = getShortMid(mid);
        String name = docs.documents[i].getField(IndexTopics.FIELD_NAME).stringValue();
        String label = docs.documents[i].getField(IndexTopics.FIELD_LABEL).stringValue();
        rankedEntities.add(new RankedEntity(shortMid, score, name, label));
    }
    return rankedEntities;
}

Example 22 with MultiFieldQueryParser

use of org.apache.lucene.queryparser.classic.MultiFieldQueryParser in project ignite by apache.

the class GridLuceneIndex method query.

/**
 * Runs lucene fulltext query over this index.
 *
 * @param qry Query.
 * @param filters Filters over result.
 * @return Query result.
 * @throws IgniteCheckedException If failed.
 */
public <K, V> GridCloseableIterator<IgniteBiTuple<K, V>> query(String qry, IndexingQueryFilter filters) throws IgniteCheckedException {
    IndexReader reader;
    try {
        long updates = updateCntr.get();
        if (updates != 0) {
            writer.commit();
            updateCntr.addAndGet(-updates);
        }
        // We can cache reader\searcher and change this to 'openIfChanged'
        reader = DirectoryReader.open(writer, true);
    } catch (IOException e) {
        throw new IgniteCheckedException(e);
    }
    IndexSearcher searcher;
    TopDocs docs;
    try {
        searcher = new IndexSearcher(reader);
        MultiFieldQueryParser parser = new MultiFieldQueryParser(idxdFields, writer.getAnalyzer());
        // parser.setAllowLeadingWildcard(true);
        // Filter expired items.
        Query filter = NumericRangeQuery.newLongRange(EXPIRATION_TIME_FIELD_NAME, U.currentTimeMillis(), null, false, false);
        BooleanQuery query = new BooleanQuery.Builder().add(parser.parse(qry), BooleanClause.Occur.MUST).add(filter, BooleanClause.Occur.FILTER).build();
        docs = searcher.search(query, Integer.MAX_VALUE);
    } catch (Exception e) {
        U.closeQuiet(reader);
        throw new IgniteCheckedException(e);
    }
    IndexingQueryCacheFilter fltr = null;
    if (filters != null)
        fltr = filters.forCache(cacheName);
    return new It<>(reader, searcher, docs.scoreDocs, fltr);
}

Example 23 with MultiFieldQueryParser

use of org.apache.lucene.queryparser.classic.MultiFieldQueryParser in project jspwiki by apache.

the class LuceneSearchProvider method findPages.

/**
 *  Searches pages using a particular combination of flags.
 *
 *  @param query The query to perform in Lucene query language
 *  @param flags A set of flags
 *  @return A Collection of SearchResult instances
 *  @throws ProviderException if there is a problem with the backend
 */
public Collection findPages(String query, int flags, WikiContext wikiContext) throws ProviderException {
    IndexSearcher searcher = null;
    ArrayList<SearchResult> list = null;
    Highlighter highlighter = null;
    try {
        String[] queryfields = { LUCENE_PAGE_CONTENTS, LUCENE_PAGE_NAME, LUCENE_AUTHOR, LUCENE_ATTACHMENTS };
        QueryParser qp = new MultiFieldQueryParser(Version.LUCENE_47, queryfields, getLuceneAnalyzer());
        // QueryParser qp = new QueryParser( LUCENE_PAGE_CONTENTS, getLuceneAnalyzer() );
        Query luceneQuery = qp.parse(query);
        if ((flags & FLAG_CONTEXTS) != 0) {
            highlighter = new Highlighter(new SimpleHTMLFormatter("<span class=\"searchmatch\">", "</span>"), new SimpleHTMLEncoder(), new QueryScorer(luceneQuery));
        }
        try {
            File dir = new File(m_luceneDirectory);
            Directory luceneDir = new SimpleFSDirectory(dir, null);
            IndexReader reader = DirectoryReader.open(luceneDir);
            searcher = new IndexSearcher(reader);
        } catch (Exception ex) {
            log.info("Lucene not yet ready; indexing not started", ex);
            return null;
        }
        ScoreDoc[] hits = searcher.search(luceneQuery, MAX_SEARCH_HITS).scoreDocs;
        AuthorizationManager mgr = m_engine.getAuthorizationManager();
        list = new ArrayList<SearchResult>(hits.length);
        for (int curr = 0; curr < hits.length; curr++) {
            int docID = hits[curr].doc;
            Document doc = searcher.doc(docID);
            String pageName = doc.get(LUCENE_ID);
            WikiPage page = m_engine.getPage(pageName, WikiPageProvider.LATEST_VERSION);
            if (page != null) {
                if (page instanceof Attachment) {
                // Currently attachments don't look nice on the search-results page
                // When the search-results are cleaned up this can be enabled again.
                }
                PagePermission pp = new PagePermission(page, PagePermission.VIEW_ACTION);
                if (mgr.checkPermission(wikiContext.getWikiSession(), pp)) {
                    int score = (int) (hits[curr].score * 100);
                    // Get highlighted search contexts
                    String text = doc.get(LUCENE_PAGE_CONTENTS);
                    String[] fragments = new String[0];
                    if (text != null && highlighter != null) {
                        TokenStream tokenStream = getLuceneAnalyzer().tokenStream(LUCENE_PAGE_CONTENTS, new StringReader(text));
                        fragments = highlighter.getBestFragments(tokenStream, text, MAX_FRAGMENTS);
                    }
                    SearchResult result = new SearchResultImpl(page, score, fragments);
                    list.add(result);
                }
            } else {
                log.error("Lucene found a result page '" + pageName + "' that could not be loaded, removing from Lucene cache");
                pageRemoved(new WikiPage(m_engine, pageName));
            }
        }
    } catch (IOException e) {
        log.error("Failed during lucene search", e);
    } catch (ParseException e) {
        log.info("Broken query; cannot parse query ", e);
        throw new ProviderException("You have entered a query Lucene cannot process: " + e.getMessage());
    } catch (InvalidTokenOffsetsException e) {
        log.error("Tokens are incompatible with provided text ", e);
    } finally {
        if (searcher != null) {
            try {
                searcher.getIndexReader().close();
            } catch (IOException e) {
                log.error(e);
            }
        }
    }
    return list;
}

Example 24 with MultiFieldQueryParser

use of org.apache.lucene.queryparser.classic.MultiFieldQueryParser in project rubia-forums by flashboss.

the class ForumsSearchModuleImpl method findTopics.

@SuppressWarnings("unchecked")
public ResultPage<Topic> findTopics(SearchCriteria criteria) throws ModuleException {
    if (criteria != null) {
        try {
            EntityManager session = getSession();
            FullTextSession fullTextSession = getFullTextSession((Session) session.getDelegate());
            Builder builder = new Builder();
            String keywords = criteria.getKeywords();
            if (keywords != null && keywords.length() != 0) {
                String[] fields = null;
                Searching searching = Searching.valueOf(criteria.getSearching());
                switch(searching) {
                    case TITLE_MSG:
                        fields = new String[] { "message.text", "topic.subject" };
                        break;
                    case MSG:
                        fields = new String[] { "message.text" };
                        break;
                }
                MultiFieldQueryParser parser = new MultiFieldQueryParser(fields, new StandardAnalyzer());
                builder.add(parser.parse(keywords), MUST);
            }
            String forumId = criteria.getForum();
            if (forumId != null && forumId.length() != 0) {
                builder.add(new TermQuery(new Term("topic.forum.id", forumId)), MUST);
            }
            String categoryId = criteria.getCategory();
            if (categoryId != null && categoryId.length() != 0) {
                builder.add(new TermQuery(new Term("topic.forum.category.id", categoryId)), MUST);
            }
            String userName = criteria.getAuthor();
            if (userName != null && userName.length() != 0) {
                builder.add(new WildcardQuery(new Term("poster.userId", userName)), MUST);
            }
            String timePeriod = criteria.getTimePeriod();
            if (timePeriod != null && timePeriod.length() != 0) {
                addPostTimeQuery(builder, TimePeriod.valueOf(timePeriod));
            }
            FullTextQuery fullTextQuery = fullTextSession.createFullTextQuery(builder.build(), Post.class);
            SortOrder sortOrder = SortOrder.valueOf(criteria.getSortOrder());
            SortBy sortBy = valueOf(criteria.getSortBy());
            fullTextQuery.setSort(getSort(sortBy, sortOrder));
            fullTextQuery.setProjection("topic.id");
            LinkedHashSet<Integer> topicIds = new LinkedHashSet<Integer>();
            LinkedHashSet<Integer> topicToDispIds = new LinkedHashSet<Integer>();
            int start = criteria.getPageSize() * criteria.getPageNumber();
            int end = start + criteria.getPageSize();
            int index = 0;
            for (Object o : fullTextQuery.list()) {
                Integer id = (Integer) ((Object[]) o)[0];
                if (topicIds.add(id)) {
                    if (index >= start && index < end) {
                        topicToDispIds.add(id);
                    }
                    index++;
                }
            }
            List<Topic> topics = null;
            if (topicToDispIds.size() > 0) {
                Query q = session.createQuery("from Topic as t join fetch t.poster where t.id IN ( :topicIds )");
                q.setParameter("topicIds", topicToDispIds);
                List<Topic> results = q.getResultList();
                topics = new LinkedList<Topic>();
                for (Integer id : topicToDispIds) {
                    for (Topic topic : results) {
                        if (id.equals(topic.getId())) {
                            topics.add(topic);
                            break;
                        }
                    }
                }
            }
            ResultPage<Topic> resultPage = new ResultPage<Topic>();
            resultPage.setPage(topics);
            resultPage.setResultSize(topicIds.size());
            return resultPage;
        } catch (ParseException e) {
            return null;
        } catch (Exception e) {
            throw new ModuleException(e.getMessage(), e);
        }
    } else {
        throw new IllegalArgumentException("criteria cannot be null");
    }
}

Example 25 with MultiFieldQueryParser

use of org.apache.lucene.queryparser.classic.MultiFieldQueryParser in project rubia-forums by flashboss.

the class ForumsSearchModuleImpl method findTopics.

@SuppressWarnings("unchecked")
public ResultPage<Topic> findTopics(SearchCriteria criteria) throws ModuleException {
    if (criteria != null) {
        try {
            EntityManager session = getSession();
            FullTextSession fullTextSession = getFullTextSession((Session) session.getDelegate());
            Builder builder = new Builder();
            String keywords = criteria.getKeywords();
            if (keywords != null && keywords.length() != 0) {
                String[] fields = null;
                Searching searching = Searching.valueOf(criteria.getSearching());
                switch(searching) {
                    case TITLE_MSG:
                        fields = new String[] { "message.text", "topic.subject" };
                        break;
                    case MSG:
                        fields = new String[] { "message.text" };
                        break;
                }
                MultiFieldQueryParser parser = new MultiFieldQueryParser(fields, new StandardAnalyzer());
                builder.add(parser.parse(keywords), MUST);
            }
            String forumId = criteria.getForum();
            if (forumId != null && forumId.length() != 0) {
                builder.add(new TermQuery(new Term("topic.forum.id", forumId)), MUST);
            }
            String categoryId = criteria.getCategory();
            if (categoryId != null && categoryId.length() != 0) {
                builder.add(new TermQuery(new Term("topic.forum.category.id", categoryId)), MUST);
            }
            String userName = criteria.getAuthor();
            if (userName != null && userName.length() != 0) {
                builder.add(new WildcardQuery(new Term("poster.userId", userName)), MUST);
            }
            String timePeriod = criteria.getTimePeriod();
            if (timePeriod != null && timePeriod.length() != 0) {
                addPostTimeQuery(builder, TimePeriod.valueOf(timePeriod));
            }
            FullTextQuery fullTextQuery = fullTextSession.createFullTextQuery(builder.build(), Post.class);
            SortOrder sortOrder = SortOrder.valueOf(criteria.getSortOrder());
            SortBy sortBy = valueOf(criteria.getSortBy());
            fullTextQuery.setSort(getSort(sortBy, sortOrder));
            fullTextQuery.setProjection("topic.id");
            LinkedHashSet<Integer> topicIds = new LinkedHashSet<Integer>();
            LinkedHashSet<Integer> topicToDispIds = new LinkedHashSet<Integer>();
            int start = criteria.getPageSize() * criteria.getPageNumber();
            int end = start + criteria.getPageSize();
            int index = 0;
            for (Object o : fullTextQuery.list()) {
                Integer id = (Integer) ((Object[]) o)[0];
                if (topicIds.add(id)) {
                    if (index >= start && index < end) {
                        topicToDispIds.add(id);
                    }
                    index++;
                }
            }
            List<Topic> topics = null;
            if (topicToDispIds.size() > 0) {
                Query q = session.createQuery("from Topic as t join fetch t.poster where t.id IN ( :topicIds )");
                q.setParameter("topicIds", topicToDispIds);
                List<Topic> results = q.getResultList();
                topics = new LinkedList<Topic>();
                for (Integer id : topicToDispIds) {
                    for (Topic topic : results) {
                        if (id.equals(topic.getId())) {
                            topics.add(topic);
                            break;
                        }
                    }
                }
            }
            ResultPage<Topic> resultPage = new ResultPage<Topic>();
            resultPage.setPage(topics);
            resultPage.setResultSize(topicIds.size());
            return resultPage;
        } catch (ParseException e) {
            return null;
        } catch (Exception e) {
            throw new ModuleException(e.getMessage(), e);
        }
    } else {
        throw new IllegalArgumentException("criteria cannot be null");
    }
}